Search

Summary The growth of pathogenic mycobacteria in phagosomes of the host cell correlates with their ability to prevent phagosome maturation. The underlying molecular mechanism remains elusive. In a previous study, we have shown that Mycobacterium avium depletes the phagosome membrane of cell surface-derived glycoconjugates (de Chastellier and Thilo, Eur. J. Cell Biol. 81, 17 – 25, 2002). We now extended these quantitative observations to the major human pathogen, Mycobacterium tuberculosis (H37Rv). At increasing times after infection of mouse bone marrow-derived macrophages, cell-surface glycoconjugates were labelled enzymatically with [ 3 H]galactose. Subsequent endocytic membrane traffic resulted in a redistribution of this label from the cell surface to endocytic membranes, including phagosomes. The steady-state distribution was measured by quantitative autoradiography at the electron microscope level. Relative to early endosomes, with which phagosomes continued to fuse and rapidly exchange membrane constituents, the phagosome membrane was depleted about 3-fold, starting during infection and in the course of 9 days thereafter. These results were in quantitative agreement with our previous observations for Mycobacterium avium . For the latter case, we now showed by cell fractionation that the depletion was selective, mainly involving glycoproteins in the 110 – 210 kDa range. Together, these results indicated that pathogenic mycobacteria induced and maintained a bulk change in phagosome membrane composition that could be of special relevance for survival of pathogenic mycobacteria within phagosomes.

Dual infection of cells may divert pathogens to intracellular compartments different from those occupied in mono-infected cells. In the present studies, mouse bone marrow in vitro-derived macrophages were first infected with virulent Mycobacterium avium, which are normally singly lodged within tight phagosomes. These phagosomes do not mature; they undergo homotypic fusion with early endosomes and do not fuse with lysosomes. Seven days later, the cultures were superinfected with phase II (non-virulent) Coxiella burnetii, organisms sheltered in lysosome- (or prelysosome)-like, multi-occupancy phagosomes. The latter can attain large size and engage in efficient homo- and heterotypic fusion with other phagosomes. Cultures were fixed for transmission electron microscopy 6, 12, 24, and 48 h later. Other M. avium-infected cultures were superinfected with amastigotes of the trypanosomatid flagellate Leishmania amazonensis, which are also sheltered in lysosome- (or prelysosome)-like multi-occupancy vacuoles, and fixed at the same time periods. Chimeric phagosomes containing both M. avium and C. burnetii, were found already at 6 h and the proportion of M. avium that colocalized with C. burnetii in the same phagosomes reached over 90% after 48 h. In such phagosomes, both organisms were ultrastructurally well preserved. In contrast, colocalization of M. avium and L. amazonensis was rarely found. Speculative scenarios that could underlie the formation of chimeric phagosomes could involve delayed maturation of C. burnetii-containing phagosomes in presence of M. avium, which would allow for fusion of C. burnetii- and M. avium-containing phagosomes; the production, by C. burnetii, of molecules that upregulate the fusion of M. avium-containing phagosomes with those that contain C. burnetii; and the secretion of factors that could favour the survival of M. avium within chimeric vacuoles.

As part of their strategy for intracellular survival, mycobacteria prevent maturation of the phagosomes in which they reside inside macrophages. The molecular basis for this inhibition is only now beginning to emerge, by way of the molecular characterisation of the phagosome membrane when it encloses virulent mycobacteria. Our own work has shown that at 15 days after the phagocytic uptake of Mycobacterium avium by mouse bone marrow-derived macrophages, the phagosome membrane is depleted about 4-fold for cell surface-derived membrane glycoconjugates, labelled by exogalactosylation, in comparison to the membrane of early endosomes with which it continues to interact. Here we asked whether this depletion occurred at early or late stages after infection. We found that only about half of the depletion had occurred at about 5 hours after the beginning of phagocytic uptake, with the remainder becoming established thereafter, with a half-time of about 2.5 days. Phagosomes became depleted in relation to early endosomes with which they continued to exchange membrane constituents. Early endosomes themselves became gradually depleted by about 30% during the 15-day post-infection period. In contrast, late endosomes/lysosomes remained unchanged, with a concentration of surface-derived glycoconjugates between that of early endosomes and of phagosomes at day 15 post infection. In view of the slowness of the post-infection change of phagosome membrane composition, we proposed that this change did not play a role in preventing maturation immediately after phagosome formation, but rather correlated with the process of maintaining the phagosomes in an immature state.

Phagosomes containing live virulent mycobacteria undergo fusion with early endosomes, but they are unable to mature normally. Accordingly, they do not fuse with lysosomes. Although M. avium-containing phagosomes retain fusion and intermingling characteristics of early endosomes indefinitely, fusions with early endosomes are increasingly restricted as bacteria multiply. In addition, when endocytic tracers, such as horseradish peroxidase (HRP), are added to M. avium-infected macrophages at 1 or up to 15 days after infection, an atypical time course of acquisition of the tracer by the phagosomes is observed, i.e., a 10 to 20 min lag, instead of immediate acquisition as is typical for early endosomes (and phagosomes with early endosome characteristics). These events coincide with a marked disorganization of the actin filament network in M. avium-infected macrophages. In the present study, we have therefore addressed the following question: Do actin filaments play a role in fusion and intermingling of contents between early endosomes and immature phagosomes that undergo homotypic fusion with early endosomes? We examined the time course of acquisition of subsequently internalized endocytic marker (HRP) by early endosome-like preexisting phagosomes, i.e. 2 hour-old phagosomes with either hydrophobic latex particles, virulent or avirulent M. avium, after depolymerization of the actin filament network with cytochalasin D or after repolymerization of the actin filament network with jasplakinolide, in cases where the network had been depolymerized (macrophages infected with M. avium, at 1 or up to 7 days after infection). By direct morphological observation at the electron microscope level and by a kinetic approach, we show here that depolymerization of the actin filament network with cytochalasin D delays acquisition of HRP whereas repolymerization restores immediate acquisition of the marker. We conclude that the actin filament network is involved in fusion and intermingling of endocytic contents between early endosomes and early endosome-like phagosomes, and that disruption of this network by M. avium is the cause for the atypical acquisition of content marker by phagosomes containing these pathogenic mycobacteria.